RESEARCH ARTICLE SPECIAL COLLECTION: TRANSLATIONAL IMPACT OF RAT

The role of the dopamine D1 receptor in social cognition: studiesusing a novel genetic rat modelJudith R. Homberg1,*, Jocelien D. A. Olivier2,*,‡, Marie VandenBroeke3, Jiun Youn3, Arabella K. Ellenbroek3,Peter Karel1, Ling Shan1, Ruben van Boxtel4, Sharon Ooms1, Monique Balemans1, Jacqueline Langedijk1,Mareike Muller1, Gert Vriend5, Alexander R. Cools1,†, Edwin Cuppen4 and Bart A. Ellenbroek3

ABSTRACTSocial cognition is an endophenotype that is impaired in schizophreniaand several other (comorbid) psychiatric disorders. One of themodulators of social cognition is dopamine, but its role is not clear.The effects of dopamine are mediated through dopamine receptors,including the dopamine D1 receptor (Drd1). Because current Drd1receptor agonists are not Drd1 selective, pharmacological tools are notsufficient to delineate the role of theDrd1. Here, we describe a novel ratmodel with a genetic mutation in Drd1 in which we measured basicbehavioural phenotypes and social cognition. The I116Smutation waspredicted to render the receptor less stable. In line with thiscomputational prediction, this Drd1 mutation led to a decreasedtransmembrane insertion of Drd1, whereas Drd1 expression, asmeasured by Drd1 mRNA levels, remained unaffected. Owing todecreased transmembrane Drd1 insertion, the mutant rats displayednormal basic motoric and neurological parameters, as well aslocomotor activity and anxiety-like behaviour. However, measures ofsocial cognition like social interaction, scent marking, pup ultrasonicvocalizations and sociability, were strongly reduced in the mutant rats.This profile of the Drd1 mutant rat offers the field of neuroscience anovel genetic rat model to study a series of psychiatric disordersincluding schizophrenia, autism, depression, bipolar disorder and drugaddiction.

KEY WORDS: Dopamine D1 receptor, Mutant rat, Social cognition,Characterization, Schizophrenia

INTRODUCTIONIt is now commonly agreed that psychiatric disorders are not fullyindependent syndromes, but rather represent constellations ofsymptom clusters that partially overlap. In support of this idea,there is extensive comorbidity between psychiatric disorders. For

instance, patients with schizophrenia often show symptoms ofmajor depressive disorder, autism spectrum disorders, bipolardisorders and/or drug addiction. Intriguingly, a very recent reportanalysing structural magnetic resonance imaging (MRI) studies hasreported that schizophrenia, and several major psychiatric disordersare jointly characterized by changes in brain areas mediating socialcognition (Goodkind et al., 2015). The ‘social brain’ is indeed acommon factor in schizophrenia and its comorbid disorders (Fettet al., 2015; Hoertnagl and Hofer, 2014; Pinkham, 2014). Socialcognition can be defined as all processes that are elicited by and/ordirected towards other subjects (Kennedy and Adolphs, 2012).Increasing effort has been made to understand the intricacies of thesocial brain (Kennedy and Adolphs, 2012; Meyer-Lindenberg andTost, 2012). One of the major players is dopamine.

Research in non-human primates and rodents has shown thatamphetamine, a dopamine-releasing drug, reduces affiliative socialbehaviour (Ellenbroek et al., 1989; Schiorring, 1979). However, it hasalso been reported that pro-social behaviour (induced by playing back50 kHz ultrasonic vocalizations that elicit approach behaviour) isassociated with increased dopamine release (Willuhn et al., 2014).These conflicting data might in part be related to opposing roles ofdifferent dopamine receptors. In this study, we focus on the role of thedopamine D1 receptor (Drd1). A recent positron emissiontomography (PET) scan study in healthy volunteers showed thatDrd1 binding was positively correlated to social conformity (Plaven-Sigray et al., 2014). In line with this, studies in zebrafish have foundthat Drd1 antagonists reduced social preference (Scerbina et al.,2012). By contrast, studies in male macaque monkeys have shownthat a Drd1 antagonist actually improved social behaviour that wasdisrupted by amphetamine (Ellenbroek et al., 1989). Likewise, localinjections of a Drd1 antagonist into the nucleus accumbens of miceleads to increases in social approach behaviour in females but notmales (Campi et al., 2014). Finally, Drd1 agonists injected into thenucleus accumbens of male prairie voles prevent new pair bonding,but facilitated the maintenance of pair bonding when the agonist wasadministered after a bond was formed (Aragona et al., 2006).

Taken together, these data emphasize that, although Drd1clearly plays an important role in social cognition, its precisefunction is far from understood. The conflicting data might – atleast in part – be due to differences in the behavioural paradigmbeing used, as well as differences in species and sex. Perhaps mostimportantly, most studies investigating the role of Drd1 in socialcognition have relied on pharmacological manipulations of Drd1.Unfortunately, none of the drugs acting on Drd1 show selectivityfor Drd1 over other dopamine receptors, and several of thesedrugs also show affinity for serotonin receptors (Undieh, 2010).Therefore, in order to investigate the role of Drd1 in socialcognition as a transdiagnostic marker for schizophrenia andcomorbid psychiatric disorders, we introduce in the present paperReceived 1 February 2016; Accepted 4 May 2016

1Donders Institute for Brain, Cognition and Behaviour, Department of CognitiveNeuroscience, Radboud University Medical Centre, Nijmegen 6525 EZ,The Netherlands. 2Department of Neurobiology, Unit Behavioural Neuroscience,Groningen Institute for Evolutionary Life Sciences, University of Groningen,Groningen 9700 CC, The Netherlands. 3Victoria University of Wellington, School ofPsychology, PO Box 600, Wellington 6040, New Zealand. 4Hubrecht Institute,KNAW and University Medical Centre Utrecht, Utrecht 3584 CT, The Netherlands.5CMBI, Radboud University Nijmegen Medical Centre, Geert Grooteplein 26–28,Nijmegen 6525 GA, The Netherlands.*These authors contributed equally to this work†Deceased

‡Author for correspondence ( j.d.a.olivier@rug.nl)

J.R.H., 0000-0001-9108-7188; J.D.A.O., 0000-0003-4654-9763

This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.

1147

© 2016. Published by The Company of Biologists Ltd | Disease Models & Mechanisms (2016) 9, 1147-1158 doi:10.1242/dmm.024752

Disea

seModels&Mechan

isms

a novel rat model with a mutation in the Drd1 gene. We firstprovide evidence that the mutant Drd1 is dysfunctional.Subsequently, we measured basic behaviour and social cognitionin the rat carrying the Drd1 mutation. To assess social cognitionwe used social interaction, social approach and avoidance,olfactory-scent marking and social anxiety (maternal-separation-induced ultrasonic vocalizations) tests. Our data show that thereceptor Drd1 is crucially involved in all these aspects of socialcognition. As these social traits are also prominent inschizophrenia and comorbid disorders, we believe our rat modelto be useful for research into these disease states.

RESULTSWe treated male Wistar rats with N-ethyl-N-nitrosourea (ENU) toinduce mutagenesis (Smits et al., 2006) and identified a rat amongthe offspring with a missense mutation in the gene encoding forDrd1. The mutation involves a hydrophobic isoleucine into polarserine residue exchange (Drd1I116S) in helix III at position 116 ofthe protein. We outcrossed this mutant rat for at least fivegenerations to wild-type Wistar rats to generate a Drd1 mutant ratmodel. We characterized and tested this mutant rat model for basicbehavioural phenotypes and social cognition.

Characterization of the Drd1I116S mutant ratComputational modelling of the Drd1I116S receptorWe first performed a computational analysis and modelling of theDrd1I116S mutation. The ENU-induced mutation changed ahydrophobic isoleucine residue into a hydrophilic serine residueat position 116. The Drd1 protein belongs to the D1-like family of

dopamine receptors that couples to the intracellular G proteinG-stimulatory (Gs). As illustrated in Fig. 1, computationalmodelling predicted that, in Drd1I116S, the cytoplasmic ends ofhelix III and helix VI cannot approach each other as closely as in thewild-type Drd1, breaking contact with the highly conserved R120 atthe extracellular side of helix III that binds to G proteins. Theseparation of the cytoplasmic ends of helix III and helix VI mightopen this part of the receptor to a larger extent, meaning that thereceptor could be rendered instable.

Dopamine receptor mRNA levelsNext, to assess whether the Drd1I116S mutation affects Drd1transcript levels, we determined mRNA expression of Drd1 in thestriatum (rich in dopamine) by quantitative PCR (qPCR). We alsoassessed mRNA levels of other dopamine receptor subtypes (Drd2–Drd5), to evaluate whether a change in Drd1 would indirectly affectthe expression of genes encoding for other dopamine receptors. Wefound that the mRNA levels of Drd1 (Z=−1.389, P=0.165), Drd2(Z=−0.579, P=0.563), Drd3 (Z=−0.000, P=1.000), Drd4 (Z=−0.347, P=0.728) and Drd5 (Z=0.000, P=1.000) were unaltered inthe striatum of wild-type rats compared to Drd1I116S mutant rats(Fig. S1). Drd4 mRNA levels were very low in both wild-type andmutant rats, which corresponds to previous findings for Drd4mRNA levels in the striatum of wild-type subjects (Matsumotoet al., 1996). None of the dopamine receptors showed any differencein levels between the wild-type and Drd1 mutant genotypes,suggesting that the Drd1I116S mutation does not affect Drd1transcript stability and/or degradation, nor that of the otherdopamine receptors.

Fig. 1. Computational modelling of thewild-type Drd1and the Drd1I116S protein structure. A computationalmodel of (A) the wild-type Drd1 receptor, and the (B)Drd1I116S receptor is shown. The solid arrow indicates theposition of the I116S mutation. The purple amino acidsreflect helix VI, and the red amino acids reflect helix III.Ligand binding takes places at about two to three helicalturns above the solid arrow. The I116S mutation (locatedjust inside the membrane near the cytosolic side of helixIII) breaks a series of hydrophobic contacts with aminoacids 276 and 277 (valine and isoleucine) in helix VI andwith the highly conserved R120 at the extracellular side ofhelix III. The exact position of this arginine residue, andthe contacts between helices III and VI are important formaintaining the receptor in an inactive state. The I116Smutant disturbs the position of R120 and the interactionsbetween helices III and VI. This might lead to a repulsionof helix VI (broken arrow), an opening for G-proteinbinding (lower side of the displayed protein) andactivation and instability of the receptor.

1148

RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 1147-1158 doi:10.1242/dmm.024752

Disea

seModels&Mechan

isms

Drd1I116S receptor bindingTo measure whether the Drd1I116S mutation would affect Drd1ligand binding we used [3H]SCH23390 (an Drd1 antagonist)autoradiography, focusing on the whole brain. This methodmeasures both transmembrane and intracellular Drd1 binding,given that cells are cut at varying positions when slicing the brain.We found that [3H]SCH23390 binding was reduced by ∼20%(olfactory tubercle) to ∼50% (prefrontal cortex, subcortical regions,substantia nigra) in the Drd1I116S mutant compared to wild-type rats(Fig. 2A; Table S1). Thus, Drd1 binding is significantly reduced inDrd1I116S mutant rats. Because Drd1 mRNA levels were unaltered,and protein synthesis presumably is unaltered as well, it is mostlikely that reduced Drd1 binding reflects reduced Drd1transmembrane insertion.

Transmembrane insertion of the Drd1I116S receptorBecause receptor instability might affect transmembrane insertionof Drd1I116S, we measured Drd1I116S cell surface expression usingan in vitro system. We cloned the wild-type and Drd1 mutantreceptor with a hemagglutinin epitope (HA) tag into COS7 cells. Byusing an anti-HA antibody we could identify the cellular position ofthe Drd1 under cell impermeable (transmembrane Drd1identification) and permeable (intracellular Drd1 identification)conditions. We found that mutant Drd1 receptor expression wasreduced in ‘live’ and impermeable COS7 cells compared to wild-type Drd1 receptor expression (Fig. 2B). The ‘fixed’ andpermeabilized cells revealed that Drd1 expression in thecytoplasm was equal for the mutant and wild-type Drd1. No Drd1immunoreactivity was observed in cells that were transfected withan empty vector. The data imply that the Drd1 mutation affects Drd1stability and thereby its transmembrane insertion.

Drd1I116S receptor functionTo assess Drd1I116S function in vivo, we measured the effects of theD1 antagonist SCH23390 on motor reflexes in the paw test(Ellenbroek et al., 1987), a behavioural test known to be sensitive toD1 receptor antagonists. Given that the data were not normallydistributed, the Mann–Whitney U-test was used to assess statisticaldifferences between treatments. In wild-type rats the Drd1antagonist SCH23390 significantly increased the forelimbreaction time (FRT; Z=−2.742, P<0.05) and hindlimb reactiontime (HRT; Z=−2.872, P<0.01; Fig. S2). In addition, in Drd1I116S

mutant rats the FRTwas significantly increased (Z=−2.605, <0.05).However, no effect was found in the HRT (Z=−1.125, notsignificant). Consequently, the SCH22390-induced increase in theHRT of Drd1I116S rats was significantly lower than in wild-type rats(Z=−2.036, P<0.05). This implies that the Drd1 is less functional inDrd1I116S mutant rats, potentially due to Drd1 instability and itstransmembrane insertion.

Behavioural characterization of the Drd1I116S mutant ratsBasic behavioural assessmentsTo characterize the mutant rats at the behavioural level, we startedwith gross phenotyping using a modified SHIRPA protocol for rats(Rogers et al., 1997), focusing on motor, sensory and neurologicalfunctions. As presented in Table S2, no motor, sensory andneurological changes were found in the Drd1I116S mutant rats.However, mean bodyweight was reduced in Drd1I116S mutant ratscompared to wild-type rats. Furthermore, upon handling,vocalization was increased in Drd1I116S mutant rats. This implieschanges in the social domain, which we further elaborated in thefinal part of this study.

Fig. 2. Drd1 ligand binding and localization in wild-type and Drd1I116S

mutant rats. (A) Representative [3H]SCH23390 autoradiographs of wild-type (WT, n=3) and Drd1I116S (MUT, n=3) rats. A 20–50% reduction of Drd1binding was found in Drd1I116S mutants compared to wild-type rats, mostlikely reflecting reduced Drd1 transmembrane insertion. This experiment wasreplicated up to four times. (B) Upper panel, Drd1I116S intracellularexpression was measured in fixed and permeabilised transfected COS7cells expressing the wild-type or mutant Drd1 receptor. The same amount ofDrd1 expression was visible in the cytoplasm for the mutant and wild-typeDrd1. Lower panel, Drd1I116S transmembrane expression was measured inlive, impermeable transfected COS7 cells expressing the wild-type or mutantDrd1 receptor. Drd1 receptor expression was reduced in impermeable COS7cells expressing the Drd1I116S compared to wild-type Drd1 receptorexpression, suggesting that the Drd1 mutation affects Drd1 transmembraneinsertion. Confocal pictures represent DNA staining (DAPI; blue), anti-HAstaining of the Drd1 receptor (green), or DNA+Drd1 staining (blue and greenmerged). This experiment was executed twice. EMPTY represents cellstransfected with empty vector.

1149

RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 1147-1158 doi:10.1242/dmm.024752

Disea

seModels&Mechan

isms

Because bodyweight was significantly reduced in Drd1I116S ratswe measured food intake in their home cage. We found that regulareating measured over four subsequent days in the home cage(Fig. S3) was significantly decreased in Drd1I116S rats [t(1,6)=5.947,P<0.05], but not when corrected for bodyweight [t(1,6)=0.428, notsignificant].Given that dopamine plays a crucial role in exploratory

behaviour, we performed more sophisticated tests to assesswhether behaviour in this domain would be affected by the Drd1mutation.We observed that Drd1I116S mutant rats did not differ fromwild-type control animals in the frequency of horizontal exploratorybehaviour bouts [t(1,14)=0.551, not significant] during a 30-min testin a circular open field (Fig. 3A), although rearing frequency wasdecreased [t(1,14)=2.956, P<0.05] in the mutant rats. Self-grooming,indicative for displacement behaviour or stereotypy, was also notaffected by the Drd1 mutation [t(1,14)=1.179, not significant].Additionally, in a square open field (Fig. 3B), there was nodifference in distance moved between genotypes [t(1,14)=0.045, notsignificant]. Using the elevated plus maze, allowing themeasurement of innate anxiety, no genotype differences werefound in the time spent on the open arm [Fig. 4A; t(1,35)=1.079, notsignificant]. General activity level, as measured by the number ofclosed arm entries, was again normal in mutant rats [Fig. 4B;t(1,35)=0.159, not significant]. These tests reveal that exploratory

behaviour and anxiety, with the exception of rearing, were notaffected in the Drd1I116S mutant rats.

Assessment of social cognition in the Drd1I116S mutant ratsSocial interactionFirst, we tested the time spent on social interaction among pairs ofunfamiliar weight-matched wild-type and Drd1I116S mutant rats.Compared to wild-type rats, the Drd1 mutant rats showed aprofound reduction in social behaviour. This is illustrated in Fig. 5.Statistical analysis (multivariate ANOVA) showed a significanteffect of genotype on active social behaviour [where one rat activelyinvestigates (sniffing, grooming etc.) the other rat] (Fig. 5;F(1,10)=14.3; P<0.005), with the Drd1I116S mutant rats showingconsistently reduced active social interaction. No significantgenotype difference was found for the duration of passive socialbehaviour (rats are in close proximity but do not actively interact)(Fig. 5).

Social approach and avoidanceTo assess social approach and avoidance, rats were tested in aT-maze with cups containing no pup, a novel pup or a familiar pup.First (phase 1), rats were habituated to the T-maze. Next (phase 2)rats were allowed to explore the maze and encountered a pupcovered by a cup on one arm, and an empty cup on the other arm.After a pause, the rats were again allowed to explore the T-maze(phase 3), now encountering a novel pup beneath the previouslyempty cup. The preference of the experimental animals for theT-maze arms was measured. During habituation, we found nodifferences between the genotypes in the total time spent on each ofthe two arms and total distance travelled (data not shown). The

Fig. 3. Exploratory behaviour and self-grooming in wild-type andDrd1I116S mutant rats.Wild-type (WT) and Drd1I116S mutant rats (MUT) (n=8)were examined for (A) frequency of rearing, horizontal exploratory behaviourand self-grooming in a circular open field. Drd1I116S mutant rats displayeddecreased rearing which could be an indicator of locomotor behaviour, butexploration did not differ from wild-type control animals. Self-grooming was notaltered in Drd1I116S mutant rats suggesting that anxiety-like behaviour is notaltered. Data represent mean+s.e.m. of the frequency. (B) Distancemoved in asquare open field. No differences were found in distance moved indicating thatthe Drd1I116S mutation has no effect on locomotor behaviour. Data representmean+s.e.m. cm moved. *P<0.05 (two-tailed Student’s t-test). The experimentwas executed once.

Fig. 4. Elevated plus-maze behaviour in wild-type and Drd1I116S mutantrats. The behaviour of wild-type (WT; n=20) and Drd1I116S mutant rats(MUT; n=17) in an elevated plus maze was examined. (A) Time spent onthe open arm. No differences were found between the Drd1I116S mutant andwild-type rats, indicating no altered anxiety-like behaviour. Data representmean+s.e.m. (B) Number of closed arm entries. No differences were foundbetween the Drd1I116S mutant and wild-type rats, indicating no effect of theDrd1I116S mutation on general activity. Data were analysed using a two-tailedStudent’s t-test. The experiment was executed once.

1150

RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 1147-1158 doi:10.1242/dmm.024752

Disea

seModels&Mechan

isms

results for the second and third phase are shown in Fig. 6. Given thatthe data were normally distributed, a mixed model ANOVA withgenotype as between and zone as within subject design was used.The data show that in phase 2, although therewas a significant effectof zone [in proximity of the cup; F(1,32)=9.7, P<0.01], there was nosignificant genotype effect or interaction (P>0.9) for the time spentin each arm (Fig. 6A). However, with respect to zone entries, there

was a significant zone [Fig. 6B; F(1,32)=10.8, P<0.005] andgenotype effect [F(1,32)=5.5, P<0.05]. More importantly, therewas a significant genotype×zone interaction [F(1,32)=4.9, P<0.05].Inspection of Fig. 6 shows that Drd1I116S mutant rats hadsignificantly less visits to the zone with the pup. In phase 3, therewas a similar difference between the time spent in the zone and zonevisits. Thus, whereas no significant main effect of zone or genotypeor interaction was found with respect to time spent in the two zones(Fig. 6C; all P>0.7), with respect to zone visits, a significant maineffect of zone (F(1,32)=6.9, P<0.05) and genotype (F(1,32)=5.8,P<0.05) was seen, as well as a significant zone×genotypeinteraction (Fig. 6D; F(1,32)=5.2, P<0.05). Inspection of Fig. 6indicated that this significant interaction was due to significantpreference for the pup in WT, but not Drd1I116S mutant rats.Summarizing, the data show that the Drd1I116S mutant rats had asignificantly reduced sociability in phase 2 and decreased interest insocial novelty in phase 3.

Scent markingUsing the scent marking test, odorant social communication can bemeasured. In this test, rats explore a lemon scent (non-socialstimulus) and a female urine scent (social stimulus). The results ofthe scent marking experiments are illustrated in Fig. 7, showing thespatial distribution of the male urine scent markings around thesocial stimulus. Given that the data were normally distributed, amixed model ANOVAwas used with genotype as a between subjectand scent as a within subject factor. There was a clear difference inscentmarking between the genotypes, with thewild-type control rats

Fig. 5. Social interaction in wild-type and Drd1I116S mutant rats. The socialinteraction of wild-type (WT) and Drd1I116S mutant (MUT) rats (n=6 pairs) wasexamined. A significant reduction in active social behaviour in Drd1I116Smutantrats was found compared to wild-type rats, indicating a social deficit inDrd1I116S mutant rats. No differences were found in passive social interaction(rats within 5 cm of each other but showing no interaction). Data are expressedas mean+s.e.m. duration of active or passive social behaviour. *P<0.005(mixed multivariate analysis). The experiment was executed once.

Fig. 6. Social approach and avoidance behaviour in wild-type and Drd1I116S mutant rats. The social approach and avoidance behaviour in wild-type(WT, N=7) and Drd1I116S mutant (MUT, N=10) rats was examined. (A) Time spent on the cup with pup (social) or on the empty cup (object) in phase II. Nodifferences were found between genotypes. Data are expressed as themean (+s.e.m.) duration of time spent around the two cups. (B) Frequency of zone visits tothe cupwith pup (social) or to the empty cup (object) in phase II. TheDrd1I116Smutant rats had significantly fewer visits to the cupwith the pup. Data are expressedasmean (+s.e.m.) number of zone visits to the two cups. These data indicate that Drd1I116S mutant rats have reduced sociability in phase II. (C) Time spent on thecup with a novel pup or on the cup with a familiar pup in phase III. No differences were found between genotypes. Data are expressed as mean (+s.e.m.)duration of time spent around the two cups. (D) Frequency of zone visits to the cup with a novel pup or to the cup with a familiar pup in phase III. Drd1I116S mutantrats significantly had fewer visits to the cup with the novel pup compared to the wild-type rats. Data are expressed as mean (+s.e.m.) number of zone visits to thetwo cups. These data suggest that Drd1I116S mutant rats have a decreased interest in social novelty in phase III. *P<0.05 (mixed model ANOVA with repeatedmeasures). The experiment was executed once.

1151

RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 1147-1158 doi:10.1242/dmm.024752

Disea

seModels&Mechan

isms

showing a much more restricted pattern (Fig. 7A), as compared to amore diffuse pattern in theDrd1I116Smutant rats (Fig. 7B) around thesocial stimulus. There were no clear differences in the scent markingpatterns around the non-social stimulus (data not shown).Statistically, this led to a significant genotype×stimulus interactionfor clustering [F(1,1022)=4.2; P<0.05], with the wild-type comparedto the Drd1I116S mutant rats showing amore clustered pattern aroundthe social stimulus as compared to the non-social stimulus (Fig. 7C).A similar genotype×stimulus interaction was found for saturation.As illustrated in Fig. 7, wild-type rats also produced much moreintense markings around the social than around the non-socialstimulus, whereas the reverse was true for the Drd1I116S mutant rats.However, due to large variability within genotype groups thisdifference failed to reach significance (Fig. 7D; P=0.08).

Ultrasonic vocalizationsFinally, we measured oral communication in the rats, and focussedon pups calling for their mother. When separated from their mother,

young pups produce typical separation calls, in the range of 30 to40 kHz. Although we analysed both males and females, nodifferences were observed and therefore results from both sexeswere combined (Fig. 8). Analysis of these calls on postnatal day 14showed that, compared to wild-type rats, Drd1I116S mutant ratscalled significantly less [F(1,16)=11.9; P<0.005]. Likewise, totalduration was significantly reduced in Drd1I116S mutant rats[F(1,14)=8.9, P<0.02] whereas the average duration per call andthe average frequency per call (in kHz) was not different betweenthe different genotypes (data not shown).

DISCUSSIONHere, we characterized a novel rat model with a genetic mutation inthe dopamine D1 receptor. Although Drd1 mRNA levels were notaffected, Drd1 binding was reduced in ex vivo brain material of theDrd1 mutant rats. This likely is due to reduced cell membraneinsertion of Drd1, as shown in vitro in COS7 cells overexpressingmutant Drd1. The paw test further revealed that the Drd1 in

Fig. 7. Scentmarking in wild-type andDrd1I116Smutant rats. The scent marking behaviour in wild-type (WT, n=14) andDrd1I116Smutant (MUT, n=22) rats wasexamined. (A) Distribution of scent markings around the female urine sample by WT rats. (B) Distribution of scent markings around the female urine sample byDrd1I116S mutant (MUT) rats. Wild-type control rats showed a much more restricted pattern in scent marking around the social stimulus compared with theDrd1I116S mutant rats. Note, the female urine sample is placed in the centre (around coordinates 2500, 2500). (C) Mean saturation (+s.e.m.) of the scent markingsaround the female and the lemon scent of WT and Drd1I116S mutant (MUT) rats. The wild-type rats showed amore clustered pattern around the social stimulus ascompared to the non-social stimulus compared to themutant rats. (D) Number of intense (mean+s.e.m.) scent markings around the female and the lemon scent ofwild-type and Drd1I116S mutant rats. Wild-type rats tended to produce much more intense markings around the social than around the non-social stimulus,whereas the reverse was true for the Drd1I116S mutant rats. Together, these data suggest that Drd1I116S mutants might have reduced rewarding effects of sniffingfemale urine. *P<0.05 urine versus lemon scent; #P=0.08 urine versus lemon scent (a mixed model ANOVA with repeated measures). The experiment wasexecuted once.

1152

RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 1147-1158 doi:10.1242/dmm.024752

Disea

seModels&Mechan

isms

Drd1I116S mutant rats was less functional compared to wild-typerats. Drd1 mutant rats did not display gross anatomical changes,although their bodyweight was significantly reduced. Nodifferences were observed in general motor, sensory andneurological functions. Likewise exploratory behaviour andanxiety-like behaviour were not affected by the Drd1I116S

mutation, except for rearing. The adapted SHIRPA test revealedan increase in vocalization upon handling of the rats, hintingtowards a role of the Drd1 in the social domain. In this line, severalaspects of social cognition were significantly reduced in theDrd1I116S mutant rats.Computational modelling and analysis of the Drd1I116S mutation

suggested crucial alterations in the mutant rat model. Owing to theDrd1I116S mutation, the cytoplasmic ends of helix III and helix IVcannot approach each other as closely as in the wild-type receptor.Therefore, this part of the receptor might be opened up to a largerextent, making it more accessible to G proteins. The Drd1I116S

mutation is also located closely to the DRY motif, which isimportant for interaction with intracellular G proteins, andmutations in this motif often lead to constitutive activity (Fanelliet al., 2009). Unfortunately, we have been unable to collect evidence(e.g. Drd1 agonist-induced cAMP production) for constitutiveactivity of the receptor. A potential reason is that, due to themutation, the receptor quickly disassembles upon Drd1 agonistbinding. These changes are also predicted to lead to lowermaximum attainable activity due to decreased transmembraneinsertion of the receptor (Ringkananont et al., 2006). To confirmthis speculation Drd1 binding was measured in an ex vivo system by[3H]SCH23390 autoradiography in the whole brain. We found thatDrd1 binding was reduced by 20–50% in Drd1I116S mutant rats.However, autoradiography does not discriminate betweentransmembrane and intracellular Drd1 binding. To resolve this, weconducted an in vitro experiment in which we overexpressed wild-type and mutant Drd1 in COS7 cells, allowing us to assess thetransmembrane insertion of the wild-type and mutant Drd1. Wefound that particularly transmembrane Drd1 insertion was reduced,with limited changes in intracellular Drd1 expression. This suggeststhat the Drd1I116S mutation affects Drd1 stability and itstransmembrane insertion, whereas the intracellular Drd1expression remains unaltered. Reduced transmembrane insertionof Drd1 likely has consequences for Drd1 function. Indeed, in thepaw test, we found that Drd1I116S mutant rats were less responsive to

the selective Drd1 receptor antagonist SCH23390. Although it is notyet fully clear how the Drd1I116S mutation affects Drd1, the datasuggest that the Drd1I116S mutant rat model represents a model forreduced Drd1 function.

We used the adapted SHIRPA test for basic phenotyping of the ratmodel. This test did not reveal motor, sensory and neurologicalconsequences of the Drd1I116S mutation. We did observe asignificantly reduced bodyweight in Drd1I116S mutant rats. Giventhat Drd1 agonists stimulate the secretion of growth hormone fromthe pituitary (Bluet-Pajot et al., 1990), reduced bodyweight mightbe explained by reduced Drd1 function. Another possibility is thatreduced food intake might cause reduced bodyweight. Of note, weobserved that food intake in the home cagewas reduced in Drd1I116S

mutants, but this difference disappeared when food intake wascorrected for bodyweight. Most likely, the reduced foodconsumption reflects the lower calorific requirements of a smallerbody. Given that animals were weight-matched in the socialinteraction experiments, it is not likely that decreased bodyweightinfluenced the differences in social behaviour.

Because dopamine plays a crucial role in exploratory behaviour,we assessed this type of behaviour in the rats when exposed to acircular and square open field, as well as the elevated plus maze.Although exploratory behaviour was not affected in these novelenvironments, Drd1I116S mutant rats showed decreased rearing inthe circular open field. Reduced rearing has been linked to decreasedlocomotor activity (Görisch and Schwarting, 2006; Thiel et al.,1999). However, we did not observe a decrease in locomotoractivity in Drd1 mutant rats. In previous work, a decrease in rearingwas also related to a decrease in food foraging, which mightcorrespond to the reduced food consumption in Drd1I116S mutantrats. The absence of exploratory changes in the elevated plus mazeindicates that there were no genotype differences in anxiety-likebehaviour. This was supported by the lack of genotype differencesin self-grooming, which in novel environments can be seen as astress-related displacement behaviour. Given that the previouslyreported relationship between Drd1 and self-grooming (Wachtelet al., 1992) is almost exclusively based on the effects of Drd1agonists and antagonists, and because – without exception – allDrd1 ligands also bind to the Drd5, our results imply that Drd5 ismore involved in self-grooming than Drd1. Taken together, thesedata indicate that general exploratory activity and anxiety levels arenormal in Drd1I116S mutant rats and do not confound behaviouralmeasures in Drd1I116S mutants.

Social cognition covers a wide variety of components, many ofwhich are disturbed in schizophrenia and other psychiatricdisorders. Social withdrawal, as measured in the social interactiontest, is considered as a measure of the negative symptoms ofschizophrenia (Wilson and Koenig, 2014). Withdrawal from socialcontact might derive from a lack of desire to have social contact. Associal interactions are able to induce conditioned place preference(El Rawas et al., 2012), the social withdrawal might be explained byanhedonia, the inability to experience pleasure or reward. Thisexplanation is supported by the scent marking test, in which malerats are allowed to approach and scent female urine, and mark thescents by urination. Given that sniffing female urine is highlyrewarding for male rats (Malkesman et al., 2010), it is possible thatscent marking around female urine marks was strongly reduced inDrd1I116S rats due to an inability to experience female urine reward.Although we did not specifically investigate olfaction in theDrd1I116S rats, the fact that there was no significant differencebetween the total number of scent markings between the genotypessuggests that olfaction per se was not affected. The reduction in

Fig. 8. Comparison of ultrasonic vocalizations made by wild-type andDrd1I116Smutant rats. The number of ultrasonic vocalizations made by 7-day-old wild-type (WT) and Drd1I116S mutant (MUT) rats during a 5-min separationfrom the mother (n=8 for each genotype). Drd1I116S mutant rats calledsignificantly less frequently compared to wild-type rats, which might reflect areward deficit. Data display (mean+s.e.m.) of ultrasonic vocalizations.***P<0.005 (mixed multivariate analysis). The experiment was executed once.

1153

RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 1147-1158 doi:10.1242/dmm.024752

Disea

seModels&Mechan

isms

scent markings around the social scent (and especially the alteredpattern and reduced saturation) suggest that the effects arespecifically related to a deficit in social cognition.Most mammalian infants, including rat pups, vocalize when

isolated. Given that interactions with adult females just beforeisolation increase vocalizations (a process termed maternalpotentiation), it is thought that the vocalization reflects a markerof pup–mother social bonds (Shair, 2014). Expression of thematernal potentiation of the ultrasonic vocalization in pups ishypothesized to be related to reward processes, in part becausedopamine activity plays a regulatory role. It has been demonstratedthat activation of dopamine type-2 receptors in the nucleusaccumbens blocks maternal potentiation without alteringvocalization rate in an initial isolation (Shair, 2014). By contrast,it has been found that Drd1 agonists, but not antagonists, reduce thenumber of isolation-induced infant rat ultrasonic vocalizations(Dastur et al., 1999; Muller et al., 2009), with the latter study evenreporting a significant increase with the D1 antagonist SCH23390.Here, we demonstrate that Drd1 dysfunction reduced the totalnumber and duration of isolation-induced rat pup ultrasonicvocalizations. At present it is difficult to explain these differences,though it should be remembered that D1 agonists and antagonistsare non-selective and also affect other receptors. Moreover, whereasin the present study ultrasonic vocalizations were measured atpostnatal day 7, Muller and colleagues used 11- or 12-day-old pups.As ultrasonic vocalizations depend strongly on the age of theanimals (Scattoni et al., 2009), this might (partly) underlie thedifferences between the studies. Regardless of this, the currentfindings are in line with the results from the other social paradigmsand might reflect a reward deficit, hindering the formation of a bondwith the mother. Potentially, this could lead to less maternal care(not measured in the present study) and feeding, and aberrantdevelopment of social behaviour.Finally, sociability and social novelty as measured in the social

approach and avoidance test were affected by the Drd1I116S

mutation. Whereas no genotype differences in duration of socialapproach and avoidance were found, the Drd1I116S mutant ratsshowed a significant reduction in the frequency of visiting the areaaround the novel pup in phase 2. Likewise, in phase 3, whereas thewild-type rats visited the novel rat significantly more frequently thanthe familiar rat, this behaviour was not seen in the Drd1I116S mutantrats. Analysis of the social approach and avoidance test usuallyincludes both total time spent in the vicinity of the cylinders andtotal time sniffing the cylinder, with the latter being the moresensitive measure (Silverman et al., 2010). In this respect, it isimportant to realize that in our analysis (using Ethovision XT) welimited the analysis to the nose point, which is closely correlated tosniffing. Thus, Drd1I116S mutant rats likely showed deficits in socialsniffing. This is reminiscent of the finding of the social interactiontest, where we found a significant reduction in active, but notpassive, social behaviour in the Drd1I116S mutant rats.This study provides an initial characterization of a novel Drd1

mutant rat model, with clear endophenotypes in the social domain.Besides that the present findings raise new research questions to beaddressed in future studies, we also like to mention some otherpotential limitations of our study. Most importantly, due to ENUmutagenesis, the Drd1I116S mutant rats might bear additionalmutations that have not been characterized. However, given thatanimals were outcrossed for at least five generations, the chance ofadditional mutations occuring is reduced to <1%. The Drd1I116S

mutant rats might share phenotypic similarities with Drd1-knockoutmice, but comparison is complicated by variable findings in these

mice. For instance, it has been reported that horizontal locomotoractivity is unaltered (Drago et al., 1996), increased (Waddingtonet al., 2005) or decreased (Smith et al., 1998) in Drd1-knockoutmice. Furthermore, it has been reported that Drd1-knockout micedisplay both reduced (Cromwell et al., 1998) and increased self-grooming behaviour (Clifford and Waddington, 1998), whereas wefound no changes in self-grooming in Drd1I116S mutant rats.Potentially, differences between species (or in genetic background)can account for the behavioural differences (Waddington et al.,2005). Indeed, studies have shown that the role of Drd1 inlocomotor activity is fundamentally different between rats and mice(Thomsen et al., 2011). Moreover, our Drd1I116S mutant rat has anoutbred Wistar background, which might substantially alter Drd1epistatic effects and approach human genetic heterogeneity to alarger extent. Finally, with the current expansion of genetic tools tomanipulate the rat genome, including zinc finger nuclease,transcription activator-like effector nucleases (TALEN) andCRISPR/CAS9, ENU mutagenesis as a technology to generatemutant rat models might seem outdated (Flister et al., 2015; Parkeret al., 2014). Although these more recent technologies, unlike ENUmutagenesis, allow targeted mutations, the advantage brought aboutby ENU mutagenesis is that it induces random mutations, not onlypremature stop codons and thereby knockout rats, but alsohypothesis-free point mutations causing amino acid exchanges(Smits et al., 2006). With ENU mutagenesis, we generated theDrd1I116S mutation. The weakness is that we do not yet completelyunderstand how the mutation affects the D1 receptor. The strength isthat the Drd1I116S mutant rat displays a clear phenotype, namely adeficit in social cognition, without a complete absence of proteinfunctioning. This might be much more relevant from a translationalpoint of view, as complete knockouts in humans are rare. Suchmutant models allow us to refine the understanding of proteinconformation and function in endophenotypes of psychiatricdisorders, and possible novel genetic routes to correct these.

In conclusion, we have characterized a novel genetic rat model forthe Drd1 allowing the assessment of the role of Drd1 in theregulation of social cognition. The data suggest that reducedtransmembrane insertion of Drd1 leads to a strong impairment invarious components of social cognition. Given that rats have,compared to mice, a more extensive behavioural repertoire (Parkeret al., 2014), particularly in the social domain, the Drd1I116S mutantrat adds to our tools to advance the understanding of mechanismsunderlying schizophrenia, autism, depression, addiction, bipolardisorder and other dopamine-related psychiatric disorders. Whereaswe focussed on social cognition, this novel rat model likely also hasunprecedented value for the assessment of the role of Drd1 in otherbehavioural domains, like reward processing and decision making.

MATERIALS AND METHODSAnimalsDrd1I116S mutant rats were generated by ENU-driven target-selectedmutagenesis on an outbred Wistar background. The Drd1I116S mutant ratscarry a missense mutation in Drd1, which resulted in an isoleucine to serineexchange (Drd1I116S) in helix III of the protein (Smits et al., 2006).Experimental animals [male wild-type (WT) and homozygous mutant(MUT)] were bred by in-crosses between heterozygous Drd1I116S rats thatwere outcrossed for at least five generations. At the age of 3 weeks animalswere genotyped. We used male rats for all experiments, unless specifiedotherwise. Rats were housed at two per cage in well-controlled rooms(temperature, 21°C±2°C, relative humidity, 60%±15%, light on between07:00 and 19:00) with water and food available ad libitum, unless specifiedotherwise. Experiments were performed in separate groups of rats at adultage (10–26 weeks), between 09.00 and 16.00. Each group of rats was

1154

RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 1147-1158 doi:10.1242/dmm.024752

Disea

seModels&Mechan

isms

exposed to one test only. All experiments were conducted with the approvalof the Animal Care Committee of the Radboud University in Nijmegen andthe Victoria University of Wellington, according to the respective laws forexperimental animals. All efforts were made to minimize the amount ofanimals and their suffering. Rats were not randomized, because groups weredetermined by genotype. Ex vivo and in vitro experiments were conducted ina blinded fashion by coding the materials. In vivo experiments could not beblinded, because wild-type and Drd1I116S mutant rats can easily bediscerned visually due to the lower bodyweight of the latter.

GenotypingGenomic DNA was isolated from ear cuts that were sampled in a 96-deepwell block (2.5 ml Riplate, Ritter) and dissolved overnight at 55°C in 300 µllysis buffer (100 mM Tris-HCl pH 8.5, 200 mM NaCl, 0.2% SDS, 5 mMEDTA and 100 µg/ml freshly added proteinase K). Tissue debris was spundown for 15 min at 15,000 g and supernatant was transferred to fresh tubes.DNA was precipitated by adding an equal volume of isopropanol, mixingand centrifugation at 15,000 g at room temperature. The supernatant wasremoved by gently inverting the block and the pellets werewashed with 70%ethanol, and dissolved in 400 µl water. Genotyping was performed using theKASPar SNP Genotyping System (KBiosciences, Hoddesdon, UK) andgene-specific primers (two allele-specific oligonucleotides of ∼40 nt inlength and one common oligonucleotide of ∼20 nt in length). Briefly, aPCR was carried out using the optimal thermocycling conditions for KTaq(94°C for 15 min; 20 cycles of 94°C for 10 s, 57°C for 20 s and 72°C for40 s; GeneAmp9700, Applied Biosystems, Foster City, CA). The PCRcontained 2 μl DNA solution, 1 μl 4× reaction mix, 15 pM reverse primerand 15 pM forward primer, 0.025 μl KTaq polymerase solution and 22 mMMgCl2 in a total volume of 4 μl. Samples were analysed in a PHERAstarplate reader (BMG Labtech, Offenburg, Germany) and data were analysedusing Klustercaller software (KBiosciences). All genotypes were confirmedin an independent reaction.

Computational analysis and modellingHomology models of the inactive and active form of the Drd1 were obtainedfrom the G-protein-coupled receptor data base (GPCRDB; Vroling et al.,2011). These models were built using the YASARA software (Krieger et al.,2002) using the protocol as described in Krieger et al. (2009) and alignmentsas provided in the GPCRDB. The structure with protein data bank (PDB)identifier 1gzm was used as template for the Drd1 in the inactive state,whereas the structure with PDB identifier 3sn6 was used as template for theactive form.

qPCRTissue punching, RNA preparation and cDNA synthesisSnap-frozen brain samples from rats (eight wild-type and eight Drd1I116S

mutant rats) were partly defrosted and the striatum was bilaterally punchedusing a 1.2-mm punching needle (Paxinos and Watson, 2004). Brain tissuesamples were homogenized with 1000 μl QIAzol Lysis Reagent (QIAGENSciences, MD) and 200 μl chloroform (Merck, Darmstadt, Germany). RNAwas isolated with an RNeasy Lipid Tissue Mini Kit (Qiagen, 74804)according to the manufacturer’s protocol. RNA concentration and qualitywas determined with a Nanodrop TMND-1000 spectrophotometer (ThermoFisher Scientific). The samples were kept at −80°C until the next day, whencDNA was made. cDNA was made with an iScript cDNA synthesis Kit(Bio-Rad, Hercules, CA) according to the manufacturer’s protocol.

qPCR procedureThe quantitative PCR (qPCR) procedures have been described previously(Shan et al., 2012). In brief, qPCR was performed in a reaction volume of20 µl, using the SYBR Green PCR kit (Promega, Madison, WI) and amixture of sense and antisense primers (2 pmol/µl). Primers used for qPCRare shown in Table S3. Reactions were run in aGeneAmp 7300 thermocyclerunder the following conditions: 2 min at 50°C and 10 min at 95°C, followedby 40 cycles of 15 s at 95°C and finally 1 min at 60°C. Data were acquiredand processed automatically by the Applied Biosystems Sequence DetectionSoftware. Specificity of amplification was checked by means of meltingcurve analysis and electrophoresis of products on a 1.5% agarose gel. Sterile

water (non-template control) and omission of reverse transcriptase (non-RTcontrol) during cDNA synthesis served as negative controls.

Amplification efficiency was determined by running qPCRs on a dilutionseries of pooled cDNA from all the subjects. Resulting cycle threshold (Ct)values were plotted against the inverse log of each dilution and the slope ofthis curve was then used to calculate the efficiency as follows: efficiency(E)=10−(1/slope). The normalization factor was based upon the geometricmean of the following four reference genes selected by geNorm analysis(Vandesompele et al., 2002): tubulin-α (TUBA), tubulin-β4 (TUBB4),glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and ubiquitin C(UBC). To minimize the variation, all qPCRs were conducted in duplicate.

Quantitative autoradiographyRats were killed by decapitation, their brains rapidly removed, frozen inliquid nitrogen and stored at −80°C. Coronal sections (16 µm) were cut on acryostat microtome at −20°C, thaw-mounted onto gelatin-coated slides andstored at−20°C until use. Frozen sections were brought to room temperatureat 60 min prior to the assay. The tissue sections (at least four slices peranimal and two animals per genotype) were pre-incubated for 20 min atroom temperature in 50 mMTris-HCL buffer at pH 7.4, containing 120 mMNaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 1 mM EDTA, 10% w/vBSA, and 1 mM ascorbate. Subsequently, the sections were incubated infresh buffer containing 1 nM [3H]SCH23390 (85.0 Ci/mmol, GEHealthcare, UK) and 40 nM ketanserin (blocking 5-HT2 receptors), in thepresence or absence of unlabelled 2 µM butaclamol (nonspecific binding)for 1 h at room temperature. The slides were then rinsed in cold 50 mMTris-HCl buffer for 15 s to remove superfluous radioligand, washed (4×5 min),rapidly dipped in cold distilled water and dried under a cold stream of air.[3H]SCH23390 sections together with [3H]Microscales™ standards (GEHealthcare, UK) were opposed to a [3H]hyperfilm (GE Healthcare, UK) andmanually developed after 1 month. The [3H]hyperfilms were scanned usinga 9200 typhoon scanner (GE Healthcare, UK). The areas of interest weredetermined using the Paxinos and Watson rat brain atlas, 6th edition (http://labs.gaidi.ca/rat-brain-atlas/). For each brain area, a fixed size square orrectangle box was placed in the area of interest across all slices. Using thetyphoon scanner, the average pixel density within the boxes was measured.This was also done for the [3H]Microscales™ standards that were used in thestandard curve. The optical densities within the brain regions of interest wereconverted into fmol/mg of tissue equivalent using this standard curve. Non-specific binding was subtracted from total [3H]SCH23390 binding.

In vitro Drd1 mutant overexpression studiesWild-type and mutant Drd1 was N-terminally fused to a hemagglutininepitope (HA) tag, cloned into an expression vector pcDNA3.1 (Invitrogen)and expressed in COS7 cells. At 24 h after transfection, cells were eitherfixed with methanol or incubated in cold DHB medium (DMEM, 25 nMHEPES and 0.2% fatty acid-free BSA) on ice for 20 min followed by 1-hincubation on ice with rabbit 1:200 anti-HA antibody (ab9110, Abcam,Cambridge, UK) in DHB medium and methanol fixation. Samples wereincubated for 1 h at room temperature with blocking buffer (1%BSA in PBSwith 0.1% Tween 20). Cells were immediately fixed after the transfectionprocedure and then incubated for 1 h at room temperature with 1:200 rabbitanti-HA antibody (Abcam) in blocking buffer. All cells were washedwith PBS three times, incubated with 1:200 goat FITC-conjugated anti-rabbit-IgG antibody in blocking buffer for 1 h in the dark. After threewashes with PBS, coverslips were mounted using Vectashield with DAPI(Brunschwigchemie, Amsterdam, The Netherlands) and cells were analysedusing confocal microscopy. The cells were checked for mycoplasmacontamination, and found to be negative.

Basic behavioural characterization of the Drd1I116S mutant ratsAdapted SHIRPAThe SHIRPA protocol was developed as a quick screen for general measuresof health, motoric and neurological parameters in genetically modifiedanimals. The procedure described by Rogers et al. (1997) was modified forthe rat. Rats were placed in a Perspex jar on a grid for 5 min and evaluated forbody position, spontaneous activity, respiration, body tremor and number offecal boli. Immediately thereafter, rats were placed in a wooden box with

1155

RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 1147-1158 doi:10.1242/dmm.024752

Disea

seModels&Mechan

isms

walls (57×57×28 cm). The floor contained separated back line drawings of16 squares. The following behaviours were recorded: transfer arousal,latency time to walk to the walls, locomotor activity, eye opening, gait,particularities of the body hair, pelvic elevation, tail position, approachreaction and touch escape. The rats were then lifted by the tail and assessedfor the position at which struggling movements occurred, as well as trunkcurl, head righting reflex and clenching with the paws to the tail. At landingon the horizontal grid, the rats were evaluated for grasping the grid bar, gripstrength when dragged by the tail across the grid, body tone after fingercompression on each side, pinna, corneal andwhisker reflexes, aswell as toe-pinch withdrawal of the hind limb after squeezing with hand-held forceps.The forepaws were then placed on a horizontal bar and the ability of the ratsto hang suspended was evaluated. Furthermore parameters like skin colour,hind limb tone, abdominal tone, lacrimation, salivation and provoked bitingin response to a pair of tweezers being put in their mouth were observed. Theair-righting reflex, together with contact righting when placed inside a smallplastic tubewere evaluated followed by the postural effects. Finally, rats wereplaced on a horizontal grid, rotated toward a vertical position of 45° andassessed for negative geotaxis, and behaviours like freezing behaviour,irritability, aggression and vocalization were scored. For all rats, thebodyweight, nose-to-tail length and rectal body temperature was measured.

Food consumption in the home cageRats were housed in genotype-matched pairs and the wirebar lids containingstandard food pellets were weighed during four consecutive days. In order tocorrect for bodyweight rats were weighed on the first day of measuring asbodyweight was not expected to change during this short period.

Paw testThe paw test was developed to assess dopaminergic functioning andconsists of a Perspex platform (30×30×20 cm) containing four holes, twohindlimbs holes (diameter: 5 cm), two forelimbs holes (diameter of 4 cm)and a slit for the tail (Ellenbroek et al., 1987). The distance between the leftand right forelimb and hindlimb holes was 15 mm, and the distance betweenforelimb and hindlimb holes was 55 cm. At 30 min before testing the ratswere treated with SCH23390 (1 mg/kg bodyweight, intraperitonealinjection). The rats were placed on the platform by inserting thehindlimbs and subsequently the forelimbs in the holes. Hind limbretraction time (HRT) and forelimb retraction time (FRT) were measured,with a minimum of 1 s and a maximum of 30 s. The test was repeated at 40and 50 min after injection. The three measurements were averaged.

Novelty-induced exploratory behaviour and self-groomingRats were placed in a transparent circular open field (diameter 20 cm) for30 min, and behaviour was videotaped. Exploratory behaviour, rearing, theduration and frequency of total grooming, and face and body grooming wereanalysed using Observer (Version 3.1, Noldus Information Technology,Wageningen, The Netherlands). Additionally, another group of animals wastested in a square open field measuring 50×50×50 cm for 30 min.Locomotor activity (distance moved, cm) was monitored usingEthoVision (Version 3.1, Noldus Information Technology).

Elevated plus-maze testThe apparatus was made of grey PVC and elevated 75 cm above the floor.The four arms (50×10 cm2) formed a cross with the central platform. Awall(height: 30 cm) of non-transparent material enclosed two arms, locatedopposite to each other. Each rat was placed on the central platform facing oneof the enclosed arms and allowed to freely explore the maze for 5 min.Behaviour was scored manually using Observer 5.0 (Noldus InformationTechnology, Wageningen, The Netherlands). The time spent on the openarm of the maze was calculated as a measure of anxiety, whereas the totalnumber of closed arm entries was considered as ameasure of general activity.

Assessment of social behaviour in the Drd1I116S mutant ratsSocial interactionAdult rats were isolated for 3 days prior to the experiment to enhance thedisplay of social behaviour. On the day of the experiment, pairs of rats (ofsimilar age, sex and genotype) from different litters were placed in a

standard translucent polypropylene box (40 cm×72 cm×22.5 cm high),under dim light conditions. The beddings of dust-free wood chips from bothsubjects’ cages were combined and transferred to the experiment box inorder to reduce the novelty of the environment. The behaviour of therats was recorded for 30 min with a video camera mounted above thecage. Duration and frequency of the following behaviours was scoredoffline using Observer® (Noldus Information Technology, Wageningen,The Netherlands) (blinded for genotype): (1) active social behaviour, bothrats are in close proximity with each other and at least one of the rats isactively investigating (sniffing, grooming etc.) the other rat; (2) passivesocial behaviour, both rats are in close proximity (within ∼5 cm of eachother) but neither animal shows active social investigation; (3) non-socialbehaviour, rats are further than 5 cm away from each other.

Social approach and avoidanceSocial approach avoidance was measured in a T-maze (arms 50 cm long,20 cm wide, with 25-cm-high walls) using a three-phase paradigm. In thefirst phase (habituation), rats were placed in the T-maze and allowed toexplore the maze. After 15 min the animal was removed and the mazecleaned with 70% ethanol. Next, two small cylindrical cups (9 cmdiameter) were placed upside down at the end of each of two arms. Underone cup a juvenile rat was placed, while the other cup was empty. Theexperimental rat was placed back and was able to explore the cups for5 min. After this second phase, the rat was removed and a new juvenile ratwas placed under the previous empty cup, while the familiar pup stayed inthe previous cup. The experimental rat was placed back again and for afinal 5 min was allowed to explore the T-maze. The behaviour of theexperimental rats was measured using Ethovision XT v.9. This programallows for a detailed tracking of behaviour, detecting nose-point, centralbody-point and tail-base point separately. For this analysis, the total timethe nose-point was within a zone of 5 cm of each of the cylinders wasanalysed. In addition, the frequency of (nose-point) zone visits wasanalysed.

Scent markingScent marking was measured using a protocol very similar to that developedfor mice (Wohr et al., 2011). Briefly, adult male rats were familiarized withfemale rats (from the same genotype) by placing them together with afemale for a period of 5 min between 5 and 7 days before the experiment, toallow them to experience female scent. On the day of the experiment, ratswere allowed to habituate to a novel round open field (diameter 80 cm).After 15 min the experimental rat was removed and the open field cleaned.Two circular pieces of filter paper (diameter 30 cm) were placed in oppositequadrants of the open field, one impregnated with 30 μl of lemon scent(non-social stimulus) and one with 30 μl of fresh urine (social stimulus)from females in oestrus. Females that were in oestrus (as determined by avaginal smear; Marcondes et al., 2002) were gently held between theforelimbs. This was usually sufficient to induce urine flow. This urine wascollected in Eppendorf tubes and used within 1 h after collection.

After both filter papers were impregnated with the smells, theexperimental male rat was placed back and allowed to explore the openfield for an additional 5 min. After this period the animal was removed, thefilter papers sprayed with ninhydrin spray (which dyes amino acids) anddried overnight. The male urine scent markings left on the filter paper sheetswere analysed by an open source software (openCFU) designed to countsmall circular objects (for the details of the program, see Geissmann, 2013).To differentiate between normal micturition and scent markings (which aremuch smaller), a maximum filter size of 10 pixels was applied to the imageprocessing. The variables analysed were the number of male urine markingssurrounding the target scent, the size of each marking (radius), saturation(intensity of each marking) and clustering (the number of markings in closeproximity to another marking).

Ultrasonic vocalizations in young ratsUltrasonic vocalizations were recorded in 7-day-old male pups. Rat pupsfrom different litters were taken from their mother and placed in a smallcircular container (diameter 9 cm) with fresh bedding material on the floor.Ultrasonic vocalizations were recorded for a period of 5 min from individual

1156

RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 1147-1158 doi:10.1242/dmm.024752

Disea

seModels&Mechan

isms

pups using Ultravox XT®. The same program was used to count all callswithin the 30–50 kHz range. Both the total number of calls and totalduration were recorded.

Statistical analysesData were checked for normality and homogeneity and analysed usingStudent’s t-tests (home-cage food consumption, novelty-inducedexploratory behaviour, self-grooming and elevated plus maze), mixedmultivariate tests (social interaction and ultrasonic vocalizations), mixedmodel ANOVA with repeated measures (social approach and avoidance,scent marking), a χ-squared test (adapted SHIRPA) or a Mann–WhitneyU-test (qPCR and paw test). All data were analysed using SPSS 16.0software (LEAD technologies, Chicago, IL). No a priori power analysis wasconducted; because of the novelty of the rat model, a priori data for a poweranalysis were not available. The level of significancewas set at P<0.05. Dataare expressed as mean±s.e.m.

This article is part of a special subject collection ‘Spotlight on Rat: TranslationalImpact’, guest edited by Tim Aitman and Aron Geurts. See related articles in thiscollection at http://dmm.biologists.org/collection/rat-disease-model.

AcknowledgementsWe thank Anthonieke Middelman for excellent technical assistance.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsJ.R.H., J.D.A.O., M.V., J.Y., A.K.E., P.K., L.S., R.v.B., S.O., M.B., J.L., M.M. andG.V.conducted experiments. E.C. and B.E. generated the Drd1 mutant rat. A.R.C.critically commented on the data. J.R.H., J.D.A.O. and B.A.E. wrote the manuscript.

FundingThis research received no specific grant from any funding agency in the public,commercial or not-for-profit sectors.

Supplementary informationSupplementary information available online athttp://dmm.biologists.org/lookup/doi/10.1242/dmm.024752.supplemental

ReferencesAragona, B. J., Liu, Y., Yu, Y. J., Curtis, J. T., Detwiler, J. M., Insel, T. R. andWang, Z. (2006). Nucleus accumbens dopamine differentially mediates theformation and maintenance of monogamous pair bonds. Nat. Neurosci. 9,133-139.

Bluet-Pajot, M. T., Mounier, F., Durand, D. and Kordon, C. (1990). Involvement ofdopamine D1 receptors in the control of growth hormone secretion in the rat.J. Endocrinol. 127, 191-196.

Campi, K. L., Greenberg, G. D., Kapoor, A., Ziegler, T. E. and Trainor, B. C.(2014). Sex differences in effects of dopamine D1 receptors on social withdrawal.Neuropharmacology 77, 208-216.

Clifford, J. J. and Waddington, J. L. (1998). Heterogeneity of behavioural profilebetween three new putative selective D3 dopamine receptor antagonists using anethologically based approach. Psychopharmacology 136, 284-290.

Cromwell, H. C., Berridge, K. C., Drago, J. and Levine, M. S. (1998). Actionsequencing is impaired in D1A-deficient mutant mice. Eur. J. Neurosci. 10,2426-2432.

Dastur, F. N., McGregor, I. S. and Brown, R. E. (1999). Dopaminergic modulationof rat pup ultrasonic vocalizations. Eur. J. Pharmacol. 382, 53-67.

Drago, J., Gerfen, C. R., Westphal, H. and Steiner, H. (1996). D1 dopaminereceptor-deficient mouse: cocaine-induced regulation of immediate-early geneand substance P expression in the striatum. Neuroscience 74, 813-823.

El Rawas, R., Klement, S., Salti, A., Fritz, M., Dechant, G., Saria, A. and Zernig,G. (2012). Preventive role of social interaction for cocaine conditioned placepreference: correlation with FosB/DeltaFosB and pCREB expression in ratmesocorticolimbic areas. Front. Behav. Neurosci. 6, 8.

Ellenbroek, B. A., Peeters, B. W., Honig, W. M. and Cools, A. R. (1987). The pawtest: a behavioural paradigm for differentiating between classical and atypicalneuroleptic drugs. Psychopharmacology 93, 343-348.

Ellenbroek, B. A., Willemen, A. P. and Cools, A. R. (1989). Are antagonists ofdopamine D1 receptors drugs that attenuate both positive and negative symptomsof schizophrenia? A pilot study in Java monkeys. Neuropsychopharmacology 2,191-199.

Fanelli, F., De Benedetti, P. G., Raimondi, F. and Seeber, M. (2009).Computational modeling of intramolecular and intermolecular communication inGPCRs. Curr. Protein Pept. Sci. 10, 173-185.

Fett, A.-K. J., Shergill, S. S. and Krabbendam, L. (2015). Social neuroscience inpsychiatry: unravelling the neural mechanisms of social dysfunction. Psychol.Med. 45, 1145-1165.

Flister, M. J., Prokop, J. W., Lazar, J., Shimoyama, M., Dwinell, M. and Geurts,A. (2015). 2015 guidelines for establishing genetically modified rat models forcardiovascular research. J. Cardiovasc. Transl. Res. 8, 269-277.

Geissmann, Q. (2013). OpenCFU, a new free and open-source software to countcell colonies and other circular objects. PLoS ONE 8, e54072.

Goodkind, M., Eickhoff, S. B., Oathes, D. J., Jiang, Y., Chang, A., Jones-Hagata,L. B., Ortega, B. N., Zaiko, Y. V., Roach, E. L., Korgaonkar, M. S. et al. (2015).Identification of a common neurobiological substrate for mental illness. JAMAPsychiatry 72, 305-315.

Gorisch, J. and Schwarting, R. K. W. (2006). Wistar rats with high versus lowrearing activity differ in radial maze performance. Neurobiol. Learn Mem. 86,175-187.

Hoertnagl, C. M. and Hofer, A. (2014). Social cognition in serious mental illness.Curr. Opin. Psychiatry 27, 197-202.

Kennedy, D. P. and Adolphs, R. (2012). The social brain in psychiatric andneurological disorders. Trends Cogn. Sci. 16, 559-572.

Krieger, E., Koraimann, G. and Vriend, G. (2002). Increasing the precision ofcomparative models with YASARA NOVA–a self-parameterizing force field.Proteins 47, 393-402.

Krieger, E., Joo, K., Lee, J., Lee, J., Raman, S., Thompson, J., Tyka, M., Baker,D. andKarplus, K. (2009). Improving physical realism, stereochemistry, and side-chain accuracy in homology modeling: four approaches that performed well inCASP8. Proteins 77 Suppl. 9, 114-122.

Malkesman, O., Scattoni, M. L., Paredes, D., Tragon, T., Pearson, B., Shaltiel,G., Chen, G., Crawley, J. N. and Manji, H. K. (2010). The female urine sniffingtest: a novel approach for assessing reward-seeking behavior in rodents. Biol.Psychiatry 67, 864-871.

Marcondes, F. K., Bianchi, F. J. and Tanno, A. P. (2002). Determination of theestrous cycle phases of rats: some helpful considerations. Braz. J. Biol. 62,609-614.

Matsumoto, M., Hidaka, K., Tada, S., Tasaki, Y. and Yamaguchi, T. (1996). Lowlevels of mRNA for dopamine D4 receptor in human cerebral cortex and striatum.J. Neurochem. 66, 915-919.

Meyer-Lindenberg, A. and Tost, H. (2012). Neural mechanisms of social risk forpsychiatric disorders. Nat. Neurosci. 15, 663-668.

Muller, J. M., Moore, H., Myers, M. M. and Shair, H. N. (2009). Dopamine’s role insocial modulation of infant isolation-induced vocalization: II. Maternally modulatedinfant separation responses are regulated by D1- and D2-family dopaminereceptors. Dev. Psychobiol. 51, 158-172.

Parker, C. C., Chen, H., Flagel, S. B., Geurts, A. M., Richards, J. B., Robinson,T. E., SolbergWoods, L. C. and Palmer, A. A. (2014). Rats are the smart choice:rationale for a renewed focus on rats in behavioral genetics. Neuropharmacology76, 250-258.

Paxinos, G. and Watson, C. (2004). The Rat Brain in Stereotaxic Coordinates, 6thEdition. ISBN-9780080475158, Ebook. Academic Press.

Pinkham, A. E. (2014). Social cognition in schizophrenia. J. Clin. Psychiatry 75Suppl. 2, 14-19.

Plaven-Sigray, P., Gustavsson, P., Farde, L., Borg, J., Stenkrona, P., Nyberg, L.,Backman, L. and Cervenka, S. (2014). Dopamine D1 receptor availability isrelated to social behavior: a positron emission tomography study. Neuroimage102, 590-595.

Ringkananont, U., Van, D. J., Montanelli, L., Ugrasbul, F., Yu, Y. M.,Weiss, R. E.,Refetoff, S. and Grasberger, H. (2006). Repulsive separation of the cytoplasmicends of transmembrane helices 3 and 6 is linked to receptor activation in a novelthyrotropin receptor mutant (M626I). Mol. Endocrinol. 20, 893-903.

Rogers, D. C., Fisher, E. M. C., Brown, S. D. M., Peters, J., Hunter, A. J. andMartin, J. E. (1997). Behavioral and functional analysis of mouse phenotype:SHIRPA, a proposed protocol for comprehensive phenotype assessment.Mamm.Genome 8, 711-713.

Scattoni, M. L., Crawley, J. and Ricceri, L. E. (2009). Ultrasonic vocalizations: atool for behavioural phenotyping of mouse models of neurodevelopmentaldisorders. Neurosci. Biobehav. Rev. 33, 508-515.

Scerbina, T., Chatterjee, D. and Gerlai, R. (2012). Dopamine receptor antagonismdisrupts social preference in zebrafish: a strain comparison study. Amino Acids43, 2059-2072.

Schiorring, (1979). Changes in individual and social behavior induced byamphetamine and related compounds in monkeys and man. In Cocaine andOther Stimulants. (ed. E. H. a. K. M. Ellinwood), pp. 481-522. New York: PlenumPress.

Shair, H. N. (2014). Parental potentiation of vocalization as a marker for filial bondsin infant animals. Dev. Psychobiol. 56, 1689-1697.

Shan, L., Bossers, K., Luchetti, S., Balesar, R., Lethbridge, N., Chazot, P. L.,Bao, A.-M. and Swaab, D. F. (2012). Alterations in the histaminergic system in the

1157

RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 1147-1158 doi:10.1242/dmm.024752

Disea

seModels&Mechan

isms

substantia nigra and striatum of Parkinson’s patients: a postmortem study.Neurobiol. Aging 33, 1488.e1-1413.e13.

Silverman, J. L., Yang, M., Lord, C. and Crawley, J. N. (2010). Behaviouralphenotyping assays for mousemodels of autism. Nat. Rev. Neurosci. 11, 490-502.

Smith, D. R., Striplin, C. D., Geller, A. M., Mailman, R. B., Drago, J., Lawler, C. P.and Gallagher, M. (1998). Behavioural assessment of mice lacking D1Adopamine receptors. Neuroscience 86, 135-146.

Smits, B. M., Mudde, J. B., van de Belt, J., Verheul, M., Olivier, J., Homberg, J.,Guryev, V., Cools, A. R., Ellenbroek, B. A., Plasterk, R. H. et al. (2006).Generation of gene knockouts and mutant models in the laboratory rat by ENU-driven target-selected mutagenesis. Pharmacogenet. Genomics 16, 159-169.

Thiel, C. M., Muller, C. P., Huston, J. P. and Schwarting, R. K.W. (1999). Highversus low reactivity to a novel environment: behavioural, pharmacological andneurochemical assessments. Neuroscience 93, 243-251.

Thomsen, M., Ralph, R. J. and Caine, S. B. (2011). Psychomotor stimulation bydopamine D(1)-like but not D(2)-like agonists in most mouse strains. Exp. Clin.Psychopharmacol. 19, 342-360.

Undieh, A. S. (2010). Pharmacology of signaling induced by dopamine D(1)-likereceptor activation. Pharmacol. Ther. 128, 37-60.

Vandesompele, J., De, P. K., Pattyn, F., Poppe, B., Van, R. N., De, P. A. andSpeleman, F. (2002). Accurate normalization of real-time quantitative RT-PCRdata by geometric averaging of multiple internal control genes. Genome Biol. 3,research0034.1.

Vroling, B., Sanders, M., Baakman, C., Borrmann, A., Verhoeven, S., Klomp, J.,Oliveira, L., de, V. J. and Vriend, G. (2011). GPCRDB: information system for Gprotein-coupled receptors. Nucleic Acids Res. 39, D309-D319.

Wachtel, S. R., Brooderson, R. J. and White, F. J. (1992). Parametric andpharmacological analyses of the enhanced grooming response elicited by the D1dopamine receptor agonist SKF 38393 in the rat. Psychopharmacology 109,41-48.

Waddington, J. L., O’Tuathaigh, C., O’Sullivan, G., Tomiyama, K., Koshikawa,N. and Croke, D. T. (2005). Phenotypic studies on dopamine receptor subtypeand associated signal transduction mutants: insights and challenges from 10years at the psychopharmacology-molecular biology interface.Psychopharmacology181, 611-638.

Willuhn, I., Tose, A., Wanat, M. J., Hart, A. S., Hollon, N. G., Phillips, P. E. M.,Schwarting, R. K. W. and Wohr, M. (2014). Phasic dopamine release in thenucleus accumbens in response to pro-social 50 kHz ultrasonic vocalizations inrats. J. Neurosci. 34, 10616-10623.

Wilson, C. A. and Koenig, J. I. (2014). Social interaction and social withdrawal inrodents as readouts for investigating the negative symptoms of schizophrenia.Eur. Neuropsychopharmacol. 24, 759-773.

Wohr, M., Roullet, F. I. and Crawley, J. N. (2011). Reduced scent marking andultrasonic vocalizations in the BTBR T+tf/J mouse model of autism. Genes BrainBehav. 10, 35-43.

1158

RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 1147-1158 doi:10.1242/dmm.024752

Disea

seModels&Mechan

isms